Cooperative learning has many pedagogical benefits. However, if the cooperative learning teams become ineffective, these benefits are lost. Accordingly, this study developed a computer‐aided assessment method for identifying ineffective teams at their early stage of dysfunction by using the Mahalanobis distance metric to examine the difference between the sequential test scores of the unknown team and the test scores of a reference group of functioning teams. The effectiveness of the proposed method was verified by conducting field experiments over an 18‐week engineering course in Taiwan. Forty‐eight students were randomly assigned to cooperative learning teams. The students' learning performance was evaluated by means of unit tests and homework tests. The functioning of the cooperative teams was examined at seven different points during the course of the study. The ineffective teams were identified with quantified type I errors. It was found that some teams failed persistently. Such teams require some form of external intervention to remedy the group dynamics. The results also showed that teams can become ineffective at any stage of the cooperative learning process. Thus, continuous monitoring is required to ensure that appropriate remedial actions are taken in a timely manner.